RLHF jobs online (12 figures)

6 of 12 platforms whose own pages we read describe ranking, comparing or rating AI model responses, the core of RLHF work, read 22 October 2026.

6 of 12 platforms Describe ranking, comparing or rating responses counting the platforms whose own pages were read, 22 Oct 2026

9 of 12 Writes prompts, answers or rubrics: not stated top answer out of 2

11 of 12 Red-teams models: not stated top answer out of 2

6 of 12 Domain tracks it names, in its own words: stated platforms stating it in their own pages (quoted below)

5 of 12 Says its general work needs no experience: not stated top answer out of 4

7 of 12 Asks an entry assessment: not stated top answer out of 2

15 Pages on file 22 named, every answer dated and hash-pinned

Ranks, compares or rates model responses

not stated6 platforms
describes ranking, comparing or rating responses6 platforms

One answer per platform across 12 platforms; a platform whose pages say different things is shown as "varies".

Counts

Row	Value
not stated	6 platforms
describes ranking, comparing or rating responses	6 platforms

Writes prompts, answers or rubrics

not stated9 platforms
describes writing prompts, answers or rubrics, or rewriting responses3 platforms

Each of 12 platforms answers once.

Counts

Row	Value
not stated	9 platforms
describes writing prompts, answers or rubrics, or rewriting responses	3 platforms

Red-teams models

not stated11 platforms
describes red-teaming or probing models for weaknesses1 platform

Each of 12 platforms answers once.

Counts

Row	Value
not stated	11 platforms
describes red-teaming or probing models for weaknesses	1 platform

Domain tracks it names, in its own words

not stated6 platforms
stated6 platforms

Each of 12 platforms answers once.

Counts

Row	Value
not stated	6 platforms
stated	6 platforms

Says its general work needs no experience

not stated5 platforms
says some work needs no experience or special skills4 platforms
asks for domain expertise, a degree or professional experience2 platforms
varies1 platform

Each of 12 platforms answers once.

Counts

Row	Value
not stated	5 platforms
says some work needs no experience or special skills	4 platforms
asks for domain expertise, a degree or professional experience	2 platforms
varies	1 platform

Asks an entry assessment

not stated7 platforms
asks an entry assessment, screening or qualification task5 platforms

Each of 12 platforms answers once.

Counts

Row	Value
not stated	7 platforms
asks an entry assessment, screening or qualification task	5 platforms

The full board, one platform per row

One platform to a row. Expand an answer for its source sentence, word for word, plus the page and the read date; "not stated" means that page does not cover the question.

Platform	Ranks, compares or rates model responses	Writes prompts, answers or rubrics	Red-teams models	Domain tracks it names, in its own words	Says its general work needs no experience	Asks an entry assessment
Alignerr	not stated Not addressed in Home page	not stated Not addressed in Home page	describes red-teaming or probing models for weaknesses Probe cutting-edge models for weaknesses and help make them safer Home page	View all roles Audio Coding General STEM View all roles Audio Coding General STEM Home page	says some work needs no experience or special skills No tech background needed, just a camera and curiosity Home page	not stated Not addressed in Home page
Clickworker	not stated Not addressed in Clickworker page	not stated Not addressed in Clickworker page	not stated Not addressed in Clickworker page	not stated Not addressed in Clickworker page	not stated Not addressed in Clickworker page	not stated Not addressed in Clickworker page
CrowdGen	describes ranking, comparing or rating responses Evaluate, annotate, generate, and refine Home page	not stated Not addressed in Home page	not stated Not addressed in Home page	not stated Not addressed in Home page	asks for domain expertise, a degree or professional experience Most of our projects don’t require technical skills, just domain expertise. Home page	asks an entry assessment, screening or qualification task Most of our projects have a qualification process. Home page
DataAnnotation	describes ranking, comparing or rating responses commonly include reviewing AI-generated text, ranking responses, checking facts Home page you’ll evaluate chatbot responses for accuracy, compare AI outputs, flag factual errors FAQ	describes writing prompts, answers or rubrics, or rewriting responses like reviewing responses, checking accuracy, refining prompts, and rating outputs Home page test AI-generated images, write challenging prompts, or review code for errors FAQ	not stated Not addressed in Home page Not addressed in FAQ	Coding Generalist Law Mathematics Medicine Physics Finance Accounting Bilingual Chemistry… Not addressed in Home page Coding Generalist Law Mathematics Medicine Physics Finance Accounting Bilingual Chemistry Biology FAQ	says some work needs no experience or special skills No coding or technical background needed. Home page No coding experience is required for general projects FAQ	asks an entry assessment, screening or qualification task Not addressed in Home page You can only take the Starter Assessment once FAQ
Handshake AI	not stated Not addressed in Programme page	not stated Not addressed in Programme page	not stated Not addressed in Programme page	not stated Not addressed in Programme page	says some work needs no experience or special skills Get paid hourly for high-impact work, no matter your background. Programme page	not stated Not addressed in Programme page
Mercor	not stated Not addressed in Open roles page	not stated Not addressed in Open roles page	not stated Not addressed in Open roles page	not stated Not addressed in Open roles page	not stated Not addressed in Open roles page	asks an entry assessment, screening or qualification task Get Instant work offers 1 Resume 2 Location 3 Assessment Open roles page
Mindrift	describes ranking, comparing or rating responses Writing expert prompts, evaluating AI-generated responses, justifying your assessments Home page Compare two options, pick the better one, or rate by criteria like accuracy and helpfulness. Welcome page for Toloka members	describes writing prompts, answers or rubrics, or rewriting responses Create realistic prompts, questions, and scenarios, the kind a real user would actually ask AI. Home page Not addressed in Welcome page for Toloka members	not stated Not addressed in Home page Not addressed in Welcome page for Toloka members	We accept experts across 90+ domains from STEM and law to creative writing and… We accept experts across 90+ domains from STEM and law to creative writing and linguistics. Home page Not addressed in Welcome page for Toloka members	varies by document You need domain expertise Home page You can start on All-access tasks right away. Welcome page for Toloka members	asks an entry assessment, screening or qualification task Complete a technical assessment to demonstrate your coding skills Home page take short qualification tasks, and unlock projects suited to your expertise Welcome page for Toloka members
OneForma	describes ranking, comparing or rating responses Apply your expertise to evaluate AI, flag what it gets wrong, and help shape what comes next. Home page	not stated Not addressed in Home page	not stated Not addressed in Home page	Marketing, psychology, journalism, architecture, generalist. If you know it, we need it. Marketing, psychology, journalism, architecture, generalist. If you know it, we need it. Home page	not stated Not addressed in Home page	not stated Not addressed in Home page
Outlier	describes ranking, comparing or rating responses Write a Challenging Prompt Create Grading Rubrics Rate and Rank Answers Home page you may generate prompts, rank AI outputs, improve model responses FAQ	describes writing prompts, answers or rubrics, or rewriting responses Think of a difficult question in your field of study Home page you may generate prompts, rank AI outputs, improve model responses FAQ	not stated Not addressed in Home page Not addressed in FAQ	We depend on subject matter experts in coding, STEM, languages, and more to improve AI We depend on subject matter experts in coding, STEM, languages, and more to improve AI Home page Not addressed in FAQ	asks for domain expertise, a degree or professional experience minimum qualifications typically require undergraduate-level expertise Home page we seek experts with deep knowledge, advanced skills, and significant experience in their fields FAQ	asks an entry assessment, screening or qualification task completing skill screenings so we can match you with the right projects Home page Review project guidelines and complete assessment tasks FAQ
Prolific	describes ranking, comparing or rating responses Evaluate machine learning models and consider key questions about AI and human interaction Participants page	not stated Not addressed in Participants page	not stated Not addressed in Participants page	not stated Not addressed in Participants page	not stated Not addressed in Participants page	not stated Not addressed in Participants page
RemoExperts	not stated Not addressed in Home page	not stated Not addressed in Home page	not stated Not addressed in Home page	Cinematographers Photoshop Experts After Effects Specialists Coding Developers STEM PhDs… Cinematographers Photoshop Experts After Effects Specialists Coding Developers STEM PhDs Software Engineers Literature Writers Generalists Specialists Home page	not stated Not addressed in Home page	not stated Not addressed in Home page
Remotasks	not stated Not addressed in Home page	not stated Not addressed in Home page	not stated Not addressed in Home page	not stated Not addressed in Home page	says some work needs no experience or special skills All you need is a computer and an internet connection. Home page	not stated Not addressed in Home page

Quoted wording for all 90 answers

Alignerr, Home page, read 2026-10-22, Ranks, compares or rates model responses: Nothing on this page answers it (1,876 characters of page text, read 2026-10-22).
Alignerr, Home page, read 2026-10-22, Writes prompts, answers or rubrics: Nothing on this page answers it (1,876 characters of page text, read 2026-10-22).
Alignerr, Home page, read 2026-10-22, Red-teams models: Probe cutting-edge models for weaknesses and help make them safer
Alignerr, Home page, read 2026-10-22, Domain tracks it names, in its own words: View all roles Audio Coding General STEM
Alignerr, Home page, read 2026-10-22, Says its general work needs no experience: No tech background needed, just a camera and curiosity
Alignerr, Home page, read 2026-10-22, Asks an entry assessment: Nothing on this page answers it (1,876 characters of page text, read 2026-10-22).
Clickworker, Clickworker page, read 2026-10-08, Ranks, compares or rates model responses: Nothing on this page answers it (9,822 characters of page text, read 2026-10-08).
Clickworker, Clickworker page, read 2026-10-08, Writes prompts, answers or rubrics: Nothing on this page answers it (9,822 characters of page text, read 2026-10-08).
Clickworker, Clickworker page, read 2026-10-08, Red-teams models: Nothing on this page answers it (9,822 characters of page text, read 2026-10-08).
Clickworker, Clickworker page, read 2026-10-08, Domain tracks it names, in its own words: Nothing on this page answers it (9,822 characters of page text, read 2026-10-08).
Clickworker, Clickworker page, read 2026-10-08, Says its general work needs no experience: Nothing on this page answers it (9,822 characters of page text, read 2026-10-08).
Clickworker, Clickworker page, read 2026-10-08, Asks an entry assessment: Nothing on this page answers it (9,822 characters of page text, read 2026-10-08).
CrowdGen, Home page, read 2026-10-08, Ranks, compares or rates model responses: Evaluate, annotate, generate, and refine
CrowdGen, Home page, read 2026-10-08, Writes prompts, answers or rubrics: Nothing on this page answers it (9,294 characters of page text, read 2026-10-08).
CrowdGen, Home page, read 2026-10-08, Red-teams models: Nothing on this page answers it (9,294 characters of page text, read 2026-10-08).
CrowdGen, Home page, read 2026-10-08, Domain tracks it names, in its own words: Nothing on this page answers it (9,294 characters of page text, read 2026-10-08).
CrowdGen, Home page, read 2026-10-08, Says its general work needs no experience: Most of our projects don’t require technical skills, just domain expertise.
CrowdGen, Home page, read 2026-10-08, Asks an entry assessment: Most of our projects have a qualification process.
DataAnnotation, Home page, read 2026-10-22, Ranks, compares or rates model responses: commonly include reviewing AI-generated text, ranking responses, checking facts
DataAnnotation, Home page, read 2026-10-22, Writes prompts, answers or rubrics: like reviewing responses, checking accuracy, refining prompts, and rating outputs
DataAnnotation, Home page, read 2026-10-22, Red-teams models: Nothing on this page answers it (7,040 characters of page text, read 2026-10-22).
DataAnnotation, Home page, read 2026-10-22, Domain tracks it names, in its own words: Nothing on this page answers it (7,040 characters of page text, read 2026-10-22).
DataAnnotation, Home page, read 2026-10-22, Says its general work needs no experience: No coding or technical background needed.
DataAnnotation, Home page, read 2026-10-22, Asks an entry assessment: Nothing on this page answers it (7,040 characters of page text, read 2026-10-22).
DataAnnotation, FAQ, read 2026-10-08, Ranks, compares or rates model responses: you’ll evaluate chatbot responses for accuracy, compare AI outputs, flag factual errors
DataAnnotation, FAQ, read 2026-10-08, Writes prompts, answers or rubrics: test AI-generated images, write challenging prompts, or review code for errors
DataAnnotation, FAQ, read 2026-10-08, Red-teams models: Nothing on this page answers it (27,215 characters of page text, read 2026-10-08).
DataAnnotation, FAQ, read 2026-10-08, Domain tracks it names, in its own words: Coding Generalist Law Mathematics Medicine Physics Finance Accounting Bilingual Chemistry Biology
DataAnnotation, FAQ, read 2026-10-08, Says its general work needs no experience: No coding experience is required for general projects
DataAnnotation, FAQ, read 2026-10-08, Asks an entry assessment: You can only take the Starter Assessment once
Handshake AI, Programme page, read 2026-10-22, Ranks, compares or rates model responses: Nothing on this page answers it (9,523 characters of page text, read 2026-10-22).
Handshake AI, Programme page, read 2026-10-22, Writes prompts, answers or rubrics: Nothing on this page answers it (9,523 characters of page text, read 2026-10-22).
Handshake AI, Programme page, read 2026-10-22, Red-teams models: Nothing on this page answers it (9,523 characters of page text, read 2026-10-22).
Handshake AI, Programme page, read 2026-10-22, Domain tracks it names, in its own words: Nothing on this page answers it (9,523 characters of page text, read 2026-10-22).
Handshake AI, Programme page, read 2026-10-22, Says its general work needs no experience: Get paid hourly for high-impact work, no matter your background.
Handshake AI, Programme page, read 2026-10-22, Asks an entry assessment: Nothing on this page answers it (9,523 characters of page text, read 2026-10-22).
Mercor, Open roles page, read 2026-10-08, Ranks, compares or rates model responses: Nothing on this page answers it (3,011 characters of page text, read 2026-10-08).
Mercor, Open roles page, read 2026-10-08, Writes prompts, answers or rubrics: Nothing on this page answers it (3,011 characters of page text, read 2026-10-08).
Mercor, Open roles page, read 2026-10-08, Red-teams models: Nothing on this page answers it (3,011 characters of page text, read 2026-10-08).
Mercor, Open roles page, read 2026-10-08, Domain tracks it names, in its own words: Nothing on this page answers it (3,011 characters of page text, read 2026-10-08).
Mercor, Open roles page, read 2026-10-08, Says its general work needs no experience: Nothing on this page answers it (3,011 characters of page text, read 2026-10-08).
Mercor, Open roles page, read 2026-10-08, Asks an entry assessment: Get Instant work offers 1 Resume 2 Location 3 Assessment
Mindrift, Home page, read 2026-10-22, Ranks, compares or rates model responses: Writing expert prompts, evaluating AI-generated responses, justifying your assessments
Mindrift, Home page, read 2026-10-22, Writes prompts, answers or rubrics: Create realistic prompts, questions, and scenarios, the kind a real user would actually ask AI.
Mindrift, Home page, read 2026-10-22, Red-teams models: Nothing on this page answers it (4,004 characters of page text, read 2026-10-22).
Mindrift, Home page, read 2026-10-22, Domain tracks it names, in its own words: We accept experts across 90+ domains from STEM and law to creative writing and linguistics.
Mindrift, Home page, read 2026-10-22, Says its general work needs no experience: You need domain expertise
Mindrift, Home page, read 2026-10-22, Asks an entry assessment: Complete a technical assessment to demonstrate your coding skills
Mindrift, Welcome page for Toloka members, read 2026-10-08, Ranks, compares or rates model responses: Compare two options, pick the better one, or rate by criteria like accuracy and helpfulness.
Mindrift, Welcome page for Toloka members, read 2026-10-08, Writes prompts, answers or rubrics: Nothing on this page answers it (5,321 characters of page text, read 2026-10-08).
Mindrift, Welcome page for Toloka members, read 2026-10-08, Red-teams models: Nothing on this page answers it (5,321 characters of page text, read 2026-10-08).
Mindrift, Welcome page for Toloka members, read 2026-10-08, Domain tracks it names, in its own words: Nothing on this page answers it (5,321 characters of page text, read 2026-10-08).
Mindrift, Welcome page for Toloka members, read 2026-10-08, Says its general work needs no experience: You can start on All-access tasks right away.
Mindrift, Welcome page for Toloka members, read 2026-10-08, Asks an entry assessment: take short qualification tasks, and unlock projects suited to your expertise
OneForma, Home page, read 2026-10-22, Ranks, compares or rates model responses: Apply your expertise to evaluate AI, flag what it gets wrong, and help shape what comes next.
OneForma, Home page, read 2026-10-22, Writes prompts, answers or rubrics: Nothing on this page answers it (7,764 characters of page text, read 2026-10-22).
OneForma, Home page, read 2026-10-22, Red-teams models: Nothing on this page answers it (7,764 characters of page text, read 2026-10-22).
OneForma, Home page, read 2026-10-22, Domain tracks it names, in its own words: Marketing, psychology, journalism, architecture, generalist. If you know it, we need it.
OneForma, Home page, read 2026-10-22, Says its general work needs no experience: Nothing on this page answers it (7,764 characters of page text, read 2026-10-22).
OneForma, Home page, read 2026-10-22, Asks an entry assessment: Nothing on this page answers it (7,764 characters of page text, read 2026-10-22).
Outlier, Home page, read 2026-10-08, Ranks, compares or rates model responses: Write a Challenging Prompt Create Grading Rubrics Rate and Rank Answers
Outlier, Home page, read 2026-10-08, Writes prompts, answers or rubrics: Think of a difficult question in your field of study
Outlier, Home page, read 2026-10-08, Red-teams models: Nothing on this page answers it (7,956 characters of page text, read 2026-10-08).
Outlier, Home page, read 2026-10-08, Domain tracks it names, in its own words: We depend on subject matter experts in coding, STEM, languages, and more to improve AI
Outlier, Home page, read 2026-10-08, Says its general work needs no experience: minimum qualifications typically require undergraduate-level expertise
Outlier, Home page, read 2026-10-08, Asks an entry assessment: completing skill screenings so we can match you with the right projects
Outlier, FAQ, read 2026-10-22, Ranks, compares or rates model responses: you may generate prompts, rank AI outputs, improve model responses
Outlier, FAQ, read 2026-10-22, Writes prompts, answers or rubrics: you may generate prompts, rank AI outputs, improve model responses
Outlier, FAQ, read 2026-10-22, Red-teams models: Nothing on this page answers it (6,004 characters of page text, read 2026-10-22).
Outlier, FAQ, read 2026-10-22, Domain tracks it names, in its own words: Nothing on this page answers it (6,004 characters of page text, read 2026-10-22).
Outlier, FAQ, read 2026-10-22, Says its general work needs no experience: we seek experts with deep knowledge, advanced skills, and significant experience in their fields
Outlier, FAQ, read 2026-10-22, Asks an entry assessment: Review project guidelines and complete assessment tasks
Prolific, Participants page, read 2026-10-08, Ranks, compares or rates model responses: Evaluate machine learning models and consider key questions about AI and human interaction
Prolific, Participants page, read 2026-10-08, Writes prompts, answers or rubrics: Nothing on this page answers it (5,958 characters of page text, read 2026-10-08).
Prolific, Participants page, read 2026-10-08, Red-teams models: Nothing on this page answers it (5,958 characters of page text, read 2026-10-08).
Prolific, Participants page, read 2026-10-08, Domain tracks it names, in its own words: Nothing on this page answers it (5,958 characters of page text, read 2026-10-08).
Prolific, Participants page, read 2026-10-08, Says its general work needs no experience: Nothing on this page answers it (5,958 characters of page text, read 2026-10-08).
Prolific, Participants page, read 2026-10-08, Asks an entry assessment: Nothing on this page answers it (5,958 characters of page text, read 2026-10-08).
RemoExperts, Home page, read 2026-10-08, Ranks, compares or rates model responses: Nothing on this page answers it (2,138 characters of page text, read 2026-10-08).
RemoExperts, Home page, read 2026-10-08, Writes prompts, answers or rubrics: Nothing on this page answers it (2,138 characters of page text, read 2026-10-08).
RemoExperts, Home page, read 2026-10-08, Red-teams models: Nothing on this page answers it (2,138 characters of page text, read 2026-10-08).
RemoExperts, Home page, read 2026-10-08, Domain tracks it names, in its own words: Cinematographers Photoshop Experts After Effects Specialists Coding Developers STEM PhDs Software Engineers Literature Writers Generalists Specialists
RemoExperts, Home page, read 2026-10-08, Says its general work needs no experience: Nothing on this page answers it (2,138 characters of page text, read 2026-10-08).
RemoExperts, Home page, read 2026-10-08, Asks an entry assessment: Nothing on this page answers it (2,138 characters of page text, read 2026-10-08).
Remotasks, Home page, read 2026-10-08, Ranks, compares or rates model responses: Nothing on this page answers it (12,158 characters of page text, read 2026-10-08).
Remotasks, Home page, read 2026-10-08, Writes prompts, answers or rubrics: Nothing on this page answers it (12,158 characters of page text, read 2026-10-08).
Remotasks, Home page, read 2026-10-08, Red-teams models: Nothing on this page answers it (12,158 characters of page text, read 2026-10-08).
Remotasks, Home page, read 2026-10-08, Domain tracks it names, in its own words: Nothing on this page answers it (12,158 characters of page text, read 2026-10-08).
Remotasks, Home page, read 2026-10-08, Says its general work needs no experience: All you need is a computer and an internet connection.
Remotasks, Home page, read 2026-10-08, Asks an entry assessment: Nothing on this page answers it (12,158 characters of page text, read 2026-10-08).

Not counted on this board

7 named documents fell outside what this board reads once we had read them; their reasons follow. They are tallied, never scored as answers.

mercor.com: Mercor | Organizing human intelligence to power the AI economy: Mercor's home page sells evaluation and data services to companies; the work is described on its open roles page, read as its own source
micro1.ai: High-paying Remote AI Training Jobs | micro1: the stored page carries only micro1's navigation (581 characters); it describes no work
micro1.ai: High-paying Remote AI Training Jobs | micro1: the stored page carries only micro1's navigation and a form notice; it describes no work
ai-support.joinhandshake.com: Handshake AI Help Center: the help centre's index lists article titles only; Handshake's work is read from its programme page
afterquery.com: Expert Data for Frontier AI: AfterQuery's home page sells datasets to AI labs; it does not describe the work to a contributor
surgehq.ai: Surge AI | Human Intelligence for AGI: Surge AI's home page sells data to AI labs; it does not describe the work to a contributor
dataforce.ai: AI Home | DataForce: DataForce's home page sells annotation and RLHF services to companies; it does not describe the work to a contributor

How the figures are worked out

Each platform's own home, FAQ and work-description pages were read through estate-fetch and stored whole (sha256 on every row).

For every page, six questions were answered from its own words: ranks, compares or rates model responses, writes prompts, answers or rubrics, red-teams models, domain tracks it names, in its own words, says its general work needs no experience, asks an entry assessment.

Every stated answer carries the sentence it rests on, checked against the stored text; 'not stated' means the page read does not address the question.

A platform with several pages reads 'varies' where they disagree.

12 platforms, 15 pages.

Source list (22)

dataannotation.tech: DataAnnotation | Future-Proof Your Career With AI Training Work (from the publisher's own home, FAQ and work-description pages; reproduced for comment)
dataannotation.tech: FAQs | DataAnnotation (from the publisher's own home, FAQ and work-description pages; reproduced for comment)
outlier.ai: Train the Next Generation of AI as a Freelancer | Outlier AI (from the publisher's own home, FAQ and work-description pages; reproduced for comment)
outlier.ai: Train the Next Generation of AI as a Freelancer | Outlier AI (from the publisher's own home, FAQ and work-description pages; reproduced for comment)
alignerr.com: Alignerr | Powered by Labelbox (from the publisher's own home, FAQ and work-description pages; reproduced for comment)
work.mercor.com: Explore Opportunities | Mercor (from the publisher's own home, FAQ and work-description pages; reproduced for comment)
mercor.com: Mercor | Organizing human intelligence to power the AI economy (read, set aside)
micro1.ai: High-paying Remote AI Training Jobs | micro1 (read, set aside)
micro1.ai: High-paying Remote AI Training Jobs | micro1 (read, set aside)
prolific.com: Prolific | Participants | Prolific (from the publisher's own home, FAQ and work-description pages; reproduced for comment)
joinhandshake.com: Handshake AI Program | Paid, Remote AI Training Work for Experts and Generalists (from the publisher's own home, FAQ and work-description pages; reproduced for comment)
ai-support.joinhandshake.com: Handshake AI Help Center (read, set aside)
remotasks.com: Remotasks | Earn $USD Doing Online Tasks from Home (from the publisher's own home, FAQ and work-description pages; reproduced for comment)
oneforma.com: OneForma | Freelance Jobs & Paid Projects Worldwide (from the publisher's own home, FAQ and work-description pages; reproduced for comment)
rex.zone: RemoExperts (from the publisher's own home, FAQ and work-description pages; reproduced for comment)
afterquery.com: Expert Data for Frontier AI (read, set aside)
clickworker.com: Become a Clickworker and earn money online (from the publisher's own home, FAQ and work-description pages; reproduced for comment)
surgehq.ai: Surge AI | Human Intelligence for AGI (read, set aside)
dataforce.ai: AI Home | DataForce (read, set aside)
mindrift.ai: AI Training Jobs, Earn $15-$100+/hr from Home | Mindrift Remote Freelance AI Work 2026 (from the publisher's own home, FAQ and work-description pages; reproduced for comment)
mindrift.ai: Mindrift.ai: AI Training Jobs Without Employment Ties, Tolokers, welcome! Join as an Independent Contributor (from the publisher's own home, FAQ and work-description pages; reproduced for comment)
crowdgen.com: CrowdGen by Appen - Where expert contributors shape frontier AI (from the publisher's own home, FAQ and work-description pages; reproduced for comment)

Reference or embed these figures

6 of 12 platforms whose own pages we read describe ranking, comparing or rating AI model responses, the core of RLHF work, read 22 October 2026.

Ranks, compares or rates model responses, by platform: not stated 6 of 12, describes ranking, comparing or rating responses 6 of 12 (12 platforms, read 22 October 2026).

Writes prompts, answers or rubrics, by platform: not stated 9 of 12, describes writing prompts, answers or rubrics, or rewriting responses 3 of 12 (12 platforms, read 22 October 2026).

Red-teams models, by platform: not stated 11 of 12, describes red-teaming or probing models for weaknesses 1 of 12 (12 platforms, read 22 October 2026).

Domain tracks it names, in its own words, by platform: not stated 6 of 12, stated 6 of 12 (12 platforms, read 22 October 2026).

Says its general work needs no experience, by platform: not stated 5 of 12, says some work needs no experience or special skills 4 of 12, asks for domain expertise, a degree or professional experience 2 of 12, varies 1 of 12 (12 platforms, read 22 October 2026).

Asks an entry assessment, by platform: not stated 7 of 12, asks an entry assessment, screening or qualification task 5 of 12 (12 platforms, read 22 October 2026).

To cite: "RLHF work by platform: the tasks, tracks and entry rules", updated 2026-10-22, https://rlhfjobs.com/datasets/rlhf-task-types/.

Put this figure on your page (HTML only)

6 of 12 platforms whose own pages we read describe ranking, comparing or rating AI model responses, the core of RLHF work, read 22 October 2026.

RLHF work by platform: the tasks, tracks and entry rules, rlhfjobs.com

<figure style="margin:0;padding:16px;border:1px solid #d0d0d0;background:#ffffff"><blockquote style="margin:0 0 8px;background:#ffffff">6 of 12 platforms whose own pages we read describe ranking, comparing or rating AI model responses, the core of RLHF work, read 22 October 2026.</blockquote><figcaption style="background:#ffffff"><a href="https://rlhfjobs.com/datasets/rlhf-task-types/" style="background:#ffffff">RLHF work by platform: the tasks, tracks and entry rules, rlhfjobs.com</a></figcaption></figure>

Also as data: JSON, CSV, Markdown.

The data as a CSV (90 answers as of 2026-10-22)

Questions this data answers

What RLHF work does each platform describe, which tracks does it offer, and does it say no experience is needed?

6 of 12 platforms whose own pages we read describe ranking, comparing or rating AI model responses, the core of RLHF work, read 22 October 2026.

Email me RLHF task changes